Abstract:

The understanding of natural languages on machines is one of the ditticult topics in artificial intelligence. Based on the current theory, the formali zation of natural language understanding is defined and its axiom system is improved in this paper. In this system, the two understanding types of ‘mea  aning understanding’ and ‘logic understanding’ are unified. According to the outputs from the execution of a prototype program for understanding chap   p, ters, the correctness and validity of all theoretical results presented in this paper are verified.

Key words: （natural language understanding, machine understanding, perception）